Most network problems I get called in to fix come from decisions that looked reasonable at the time, like saving money on a backup circuit or paying for more bandwidth than a location needed. Unfortunately, those decisions get expensive during an outage: Uptime’s 2025 Annual Outage Analysis found that 54% of organizations say their most recent serious outage cost them more than $100,000.1
After nineteen years advising mid-market IT leaders, I see the same three patterns again and again, and I'd bet you’ll recognize at least one.
None of these mistakes come from carelessness. They usually come from treating connectivity as a line item instead of a business decision, and each one tends to stay hidden until an outage exposes it.
The most common mistake I see is a location running on a single circuit with nothing behind it, so when that circuit goes down, the business goes down too.
Internet costs have come way down, so every location should explore redundancy. A low-cost broadband circuit that can fail over to a dedicated fiber circuit (or another broadband circuit) is ideal, and the options keep getting better:
Adding one of these options is usually a small monthly cost compared with what an outage can do to a location that has no backup. The right mix depends on what’s available at each site and how much traffic the backup needs to carry.
The second mistake is buying dedicated internet access (DIA) for every location by default. DIA is the preferred and most reliable option, and if your budget can afford it, do it.
The problem is price. In most markets I work in, 100 Mbps of DIA runs about $500 a month and 1 Gbps runs about $1,000. Multiply that across a dozen branch offices and the math gets hard to defend for sites that would run well on broadband.
A better design for smaller offices looks like this:
That combination gives you diversity at a fraction of the cost. Explore what each of these connection types does in our guide to modern internet architectures.
Plenty of companies assume they’re covered with a backup circuit, but they’ve never actually confirmed it can carry the business. Low-cost broadband or cellular backup often can’t handle production traffic once failover kicks in, so the site stays technically online while the apps users depend on stop working.
I recommend testing failover every six months, because traffic grows and a backup sized two years ago may not be enough today.
Once you’ve spotted one of these mistakes, you might consider ripping everything out and starting over. However, the cost alone makes that option unrealistic.
Here’s a smarter approach:
The fix is usually to add the right secondary circuit or right-size your primary, which is significantly cheaper than replacing your entire network.
If any of these mistakes sound familiar, let’s have a conversation before your next outage forces you to. OnDemand’s team of network experts can review your locations for redundancy and testing gaps and recommend options that fit your budget – all at no cost to you.
Contact us today to start a no-obligation network strategy conversation.
Sources: